# Dynamical Systems (math.DS)

• We propose a deep generative Markov State Model (DeepGenMSM) learning framework for inference of metastable dynamical systems and prediction of trajectories. After unsupervised training on time series data, the model contains (i) a probabilistic encoder that maps from high-dimensional configuration space to a small-sized vector indicating the membership to metastable (long-lived) states, (ii) a Markov chain that governs the transitions between metastable states and facilitates analysis of the long-time dynamics, and (iii) a generative part that samples the conditional distribution of configurations in the next time step. The model can be operated in a recursive fashion to generate trajectories to predict the system evolution from a defined starting state and propose new configurations. The DeepGenMSM is demonstrated to provide accurate estimates of the long-time kinetics and generate valid distributions for molecular dynamics (MD) benchmark systems. Remarkably, we show that DeepGenMSMs are able to make long time-steps in molecular configuration space and generate physically realistic structures in regions that were not seen in training data.
• For one parameter subgroup action on a finite volume homogeneous space, we consider the set of points admitting divergent on average trajectories. We show that the Hausdorff dimension of this set is strictly less than the manifold dimension of the homogeneous space. As a corollary we know that the Hausdorff dimension of the set of points admitting divergent trajectories is not full, which proves a conjecture of Y. Cheung.
• A major challenge in the study of dynamical systems is that of model discovery: turning data into models that are not just predictive, but provide insight into the nature of the underlying dynamical system that generated the data. This problem is made more difficult by the fact that many systems of interest exhibit diverse behaviors across multiple time scales. We introduce a number of data-driven strategies for discovering nonlinear multiscale dynamical systems and their embeddings from data. We consider two canonical cases: (i) systems for which we have full measurements of the governing variables, and (ii) systems for which we have incomplete measurements. For systems with full state measurements, we show that the recent sparse identification of nonlinear dynamical systems (SINDy) method can discover governing equations with relatively little data and introduce a sampling method that allows SINDy to scale efficiently to problems with multiple time scales. Specifically, we can discover distinct governing equations at slow and fast scales. For systems with incomplete observations, we show that the Hankel alternative view of Koopman (HAVOK) method, based on time-delay embedding coordinates, can be used to obtain a linear model and Koopman invariant measurement system that nearly perfectly captures the dynamics of nonlinear quasiperiodic systems. We introduce two strategies for using HAVOK on systems with multiple time scales. Together, our approaches provide a suite of mathematical strategies for reducing the data required to discover and model nonlinear multiscale systems.
• May 22 2018 math.DS arXiv:1805.08050v1
We consider random iteration of exponential entire functions, i.e. of the form ${\mathbb C}\ni z\mapsto f_\lambda(z):=\lambda e^z\in\mathbb C$, $\lambda\in{\mathbb C}\setminus \{0\}$. Assuming that $\lambda$ is in a bounded closed interval $[A,B]$ with $A>1/e$, we deal with random iteration of the maps $f_\lambda$ governed by an invertible measurable map $\theta:\Omega\to\Omega$ preserving a probability ergodic measure $m$ on $\Omega$, where $\Omega$ is a measurable space. The link from $\Omega$ to exponential maps is then given by an arbitrary measurable function $\eta:\Omega\longmapsto [A,B]$. We in fact work on the cylinder space $Q:={\mathbb C}/\sim$, where $\sim$ is the natural equivalence relation: $z\sim w$ if and only if $w-z$ is an integral multiple of $2\pi i$. We prove that then for every $t>1$ there exists a unique random conformal measure $\nu^{(t)}$ for the random conformal dynamical system on $Q$. We further prove that this measure is supported on the, appropriately defined, radial Julia set. Next, we show that there exists a unique random probability invariant measure $\mu^{(t)}$ absolutely continuous with respect to $\mu^{(t)}$. In fact $\mu^{(t)}$ is equivalent with $\nu^{(t)}$. Then we turn to geometry. We define an expected topological pressure $\mathcal E P(t)\in{\mathbb R}$ and show that its only zero $h$ coincides with the Hausdorff dimension of $m$--almost every fiber radial Julia set $J_r(\omega)\subset Q$, $\omega\in\Omega$. We show that $h\in (1,2)$ and that the omega--limit set of Lebesgue almost every point in $Q$ is contained in the real line $\mathbb R$. Finally, we entirely transfer our results to the original random dynamical system on $\mathbb C$. As our preliminary result, we show that all fiber Julia sets coincide with the entire complex plane $\mathbb C$.
• In structural dynamics, energy dissipative mechanisms with non-viscous damping are characterized by their dependence on the time-history of the response velocity, mathematically represented by convolution integrals involving hereditary functions. Combination of damping parameters in the dissipative model can lead the system to be overdamped in some (or all) modes. In the domain of the damping parameters, the thresholds between induced oscillatory and non--oscillatory motion are called critical damping surfaces (or manifolds, since we can have a lot of parameters). In this paper a general method to obtain critical damping surfaces for nonviscously damped systems is proposed. The approach is based on transforming the algebraic equations which defined implicitly the critical curves into a system of differential equations. The derivations are validated with three numerical methods covering single and multiple degree of freedom systems.
• Working on strongly irreducible planar self-affine sets satisfying the strong open set condition, we calculate the Birkhoff spectrum of continuous potentials and the Lyapunov spectrum.
• In this paper, we study the large scaled geometric structure of Julia sets of entire and meromorphic functions. Roughly speaking, the structure gives us some asymptotic information about the Julia set near the essential singularity. We will show that part of this structure are determined by the transcendental directions coming from function theoretic point of view.
• The paper considers two-dimensional spectral submanifolds (SSM) of equilibria of finite dimensional vector fields. SSMs are the smoothest invariant manifolds tangent to an invariant linear subspace of an equilibrium. The paper assumes that the vector field becomes conservative at the zero limit of a parameter. It is known that in the conservative limit there exists a unique sub-centre manifold. It is also known that the non-conservative system has a unique SSM under some conditions. However, it is not clear whether the sub-centre manifold is the limit of the SSM and if this limit is smoothly approached. Here, we show that the unique SSM smoothly approaches the sub-centre manifold as the system tends to the conservative limit. We also introduce a numerical scheme to calculate the SSM or the sub-centre manifold. We find that the error of calculating the SSM grows polynomially with the distance from the equilibrium. The exponent of growth of the error strictly depends on the eigenvalues of the Jacobian matrix about the equilibrium. We argue that such polynomial growth of error is unavoidable by numerical schemes.
• This paper is concerned with pullback dynamics of a 3D Navier-Stokes equations with variable viscosity and subject to perturbations of time-dependent external forces. Under suitable assumptions on the external force, which is possibly unbounded, we establish the existence of finite-dimensional minimal pullback attractor in a general setting involving tempered universe. We also present a sufficient condition on the viscosity coefficients in order for the attractors to be non-trivial. We conclude the paper by showing the upper semi-continuity of pullback attractors as the non-autonomous perturbation vanishes.
• We first obtain a locally uniform a priori bound of the dynamics of rational functions of degree $>1$ on the Berkovich projective line over an algebraically closed and complete non-archimedean field of any characteristic, and an equidistribution result for moving targets towards the equilibrium (or canonical) measure, under the no potentially good reductions condition. This answers a question posed by Favre and Rivera-Letelier. We then establish a complex counterpart to the above a priori bound, on the dynamics of an endomorphism $f$ of $\mathbb{C}\mathbb{P}^k$ of degree $>1$. As a special case, this yields a Diophantine-type estimate of the dynamics of $f$ on its domaines singuliers (rotation domains).
• Let $(X,\mathcal{A}, \mu)$ be a probability measure space and let $T_i,$ $1\leq i\leq H,$ be invertible bi measurable measure preserving transformations on this measure space. We give a sufficient condition for the product of $H$ bounded functions $f_1, f_2, ..., f_H$ to be a coboundary. This condition turns out to be also necessary when one seeks bounded coboundaries.
• We present a stability analysis of the standard nonautonomous systems type for a recently introduced generalized Lane-Emden equation which is shown to explain the presence of some of the structures observed in the atomic spatial distributions of magnetically-trapped ultracold atomic clouds. A Lyapunov function is defined which helps us to prove that stable spatial structures in the atomic clouds exist only for adiabatic index $\gamma=1+1/n$ with even $n$. In the case when $n$ is odd we provide an instability result indicating the divergence of the density function for the atoms. Several numerical solutions, which according to our stability analysis are stable, are also presented.
• In this paper, the interrelations of some dynamical properties of the non-autonomous dynamical system (X, f1;infinity) and its induced non-autonomous dynamical system (K(X), f1;infinity) are studied, where K(X) is the hyperspace of all non-empty compact subsets of X, endowed with Vietoris topology. Various stronger forms of sensitivity and transitivity are considered. Some examples of non-autonomous systems are provided to support the results. A relation between shadowing property of the non- autonomous system (X, f1;infinity) and its induced system (K(X), f1;infinity) is studied.
• We define some pointwise properties of topological dynamical systems and give pointwise conditions for such a system possesses positive topological entropy. We give sufficient conditions to obtain positive topological entropy for maps which are approximated by maps with the shadowing property in a uniform way.
• We introduce an algorithm to locate contours of functions that are expensive to evaluate. The problem of locating contours arise in many applications, including classification, constrained optimization, and analysis of performance of mechanical and dynamical systems (reliability, probability of failure, stability, etc.). Our algorithm locates contours using information from multiple sources, which are available in the form of relatively inexpensive, biased, and possibly noisy approximations to the original function. Considering multiple information sources can lead to significant cost savings. We also introduce the concept of contour entropy, a formal measure of uncertainty about the location of the zero contour of a function approximated by a statistical surrogate model. Our algorithm locates contours efficiently by maximizing the reduction of contour entropy per unit cost.
• We give an elementary and short proof of the classical pointwise ergodic theorem for probability measure preserving $\mathbb{Z}$-actions. The main idea is to tightly tile each orbit by intervals with desired averages.
• Delay-coordinate embedding is a powerful, time-tested mathematical framework for reconstructing the dynamics of a system from a series of scalar observations. Most of the associated theory and heuristics are overly stringent for real-world data, however, and real-time use is out of the question due to the expert human intuition needed to use these heuristics correctly. The approach outlined in this thesis represents a paradigm shift away from that traditional approach. I argue that perfect reconstructions are not only unnecessary for the purposes of delay-coordinate based forecasting, but that they can often be less effective than reduced-order versions of those same models. I demonstrate this using a range of low- and high-dimensional dynamical systems, showing that forecast models that employ imperfect reconstructions of the dynamics---i.e., models that are not necessarily true embeddings---can produce surprisingly accurate predictions of the future state of these systems. I develop a theoretical framework for understanding why this is so. This framework, which combines information theory and computational topology, also allows one to quantify the amount of predictive structure in a given time series, and even to choose which forecast method will be the most effective for those data.